It is a well-established empirical regularity that union members are more dissatisfied with their jobs than their non-union counterparts. The original theoretical explanation for this seeming paradox is that unions make workers dissatisfied with their terms and conditions of employment in order to galvanise them into pressing for better conditions. However, this theory has been challenged by alternative explanations which posit that union member dissatisfaction may be the result of the sorting of dissatisfied workers into union membership and/or poorer working conditions in the union sector which prompt unionisation. The novel contribution of this paper is to test the theory that it is union membership (as opposed to sorting or working conditions) that causes job dissatisfaction using quasi-experimental methods. The article utilises difference-in-differences analysis to measure the change in job satisfaction associated with an exogenous change in working conditions that affected some union members and non-members (under open-shop arrangements) equally, comparing this to changes in job satisfaction for union and non-union members who were not affected by the exogenous change. The exogenous change in question is a proposed change in the pension arrangements for 5 million UK public sector workers that was announced in March 2011. The proposals triggered a campaign of union resistance involving co-ordinated action by 30 separate unions representing 2.1 million members. The study draws on the United Kingdom Household Longitudinal Study (UKHLS). It exploits the fact that the dispute took place between waves of this survey. Results suggest that unions do not cause job dissatisfaction.